results/PatchTST/config.json

{
    "model": "PatchTST",
    "model_type": "deep-learning",
    "model_dtype": "float32"
}